--- title: "Base Rate" description: "A base rate is how common something is in the population before you look at any specific evidence. Base-rate neglect happens when vivid individual detail displaces that background frequency, which is why a positive test for a rare condition usually still means you do not have it." last_updated: "2026-08-02" --- # Base Rate A base rate is how common something is in the population before you look at any specific evidence. Base-rate neglect happens when vivid individual detail displaces that background frequency, which is why a positive test for a rare condition usually still means you do not have it. The underlying prevalence of something in a population. Ignoring it in favor of vivid specific detail leads to base-rate neglect and miscalibrated probabilities. *Example: A test for a 1-in-10,000 disease flags you positive — but because the disease is so rare, you're still probably fine.* ## What it looks like **In a medical screening result** > The test is 99% accurate and you tested positive, so you almost certainly have the disease. *Why it fails:* If the condition affects 1 in 10,000 people, most positives among 10,000 tested are false positives — roughly 100 of them against a single true case. Accuracy alone cannot tell you what a positive means. **In a hiring conversation** > He seems quiet and reads a lot, so he's more likely a librarian than a salesperson. *Why it fails:* There are vastly more salespeople than librarians, so even a strong personality match is swamped by the difference in how many of each exist to begin with. **In a security briefing** > The system flagged this transaction as fraudulent, so we should freeze the account. *Why it fails:* With fraud rare and transactions numerous, even a low false-positive rate produces far more flagged legitimate transactions than fraudulent ones. ## Practice ### Question 1 A disease affects 1 in 1,000 people. A test has no false negatives and a 5% false-positive rate. You test positive. Roughly what is the chance you have the disease? A. About 95% B. About 50% C. About 2% D. About 80% **Answer: C.** Take 1,000 people: 1 has the disease and tests positive; of the 999 healthy, about 50 test positive falsely. So roughly 51 positives, of which 1 is real — about 2%. Most people answer 95% by reading the test's accuracy as the answer, which ignores how rare the condition is. This is the classic demonstration of base-rate neglect. ### Question 2 Why does a highly accurate test still produce mostly false positives for rare conditions? A. Because the tests are poorly manufactured B. Because the healthy group is so much larger that even a small error rate applied to it outnumbers the true cases in the small sick group C. Because rare diseases are harder to detect D. Because accuracy figures are usually exaggerated **Answer: B.** It is arithmetic about group sizes, not a defect in the test. A 5% error rate applied to 999 healthy people yields about 50 false positives; a perfect detection rate applied to 1 sick person yields 1 true positive. The larger group's small error rate dominates the smaller group's perfect one. ### Question 3 Which framing makes base-rate problems easiest for most people to solve correctly? A. Stating the probabilities as percentages B. Stating them as natural frequencies — 'out of 1,000 people, 1 has it and 50 test positive falsely' C. Stating them as decimal odds D. Presenting the formula first **Answer: B.** Gigerenzer's work found that natural-frequency framing substantially improves performance on exactly these problems, including among physicians. The information is identical; counting people rather than manipulating percentages makes the relative group sizes visible instead of leaving them implicit. ### Question 4 A witness identifies a taxi as blue. The witness is 80% reliable, but 85% of taxis in the city are green. What should you conclude? A. The taxi was probably blue, since the witness is reliable B. The taxi was more likely green — the strong prior against blue outweighs an 80% reliable identification C. The witness testimony is worthless D. There is no way to reason about this **Answer: B.** Working it through: of 100 taxis, 15 blue and 85 green. The witness correctly identifies 12 of the blue ones and wrongly calls 17 of the green ones blue — so 'blue' identifications are mostly wrong. The testimony is genuine evidence, which is why option C is wrong; it is simply not strong enough to overcome the prior. ### Question 5 What is the practical habit that guards against base-rate neglect? A. Distrust all statistics B. Ask how common the thing is in general before weighing the specific evidence in front of you C. Always assume the rarer option D. Require a second test before believing any result **Answer: B.** Establishing the prior first anchors the evidence appropriately, because the same evidence means very different things against different base rates. Option C inverts the error rather than fixing it. Option D is sometimes good practice but does not address the reasoning failure — it just adds data without changing how you weigh it. ## Frequently asked ### What is base-rate neglect? Ignoring how common something is in the population when judging a specific case, so vivid individual detail displaces background frequency. It is why people overestimate the meaning of a positive test for a rare condition and why unusual explanations feel more probable than they are. ### How is base rate related to Bayesian reasoning? The base rate is the prior probability that Bayesian updating starts from. Bayes' rule is the formal method for combining that prior with new evidence; base-rate neglect is what happens when the prior is dropped from the calculation entirely and the evidence is read on its own. ### Does this only matter in medicine? No. It applies anywhere a screening process searches a large population for something rare — fraud detection, security screening, hiring filters, and spam classification all have the same structure. It also affects everyday judgements about how likely an unusual explanation is. ## Related - [Bayesian Reasoning](https://www.criticalthinkingexercise.org/pages/glossary/bayesian-reasoning.md) - [Availability Heuristic](https://www.criticalthinkingexercise.org/pages/glossary/availability-heuristic.md) - [Hasty Generalization](https://www.criticalthinkingexercise.org/pages/glossary/hasty-generalization.md) - [Basic Probability Intuition](https://www.criticalthinkingexercise.org/pages/exercises/probability-intuition.md) - [Statistical Fallacies](https://www.criticalthinkingexercise.org/pages/exercises/statistical-fallacies.md) - [Lesson: traps-and-pitfalls/cognitive-biases-decision-social](https://www.criticalthinkingexercise.org/pages/learn/traps-and-pitfalls/cognitive-biases-decision-social.md) ## Sitemap See the full [sitemap](https://www.criticalthinkingexercise.org/sitemap.md) for all pages.